What problem does it solve? Raw clinical trial data from EDC systems and CRFs often contains missing values, inconsistent date formats, and outliers that block FDA or EMA regulatory submissions. This Skill automates the cleaning and standardization of that data into CDISC SDTM-compliant datasets while generating a complete audit trail for compliance. ## Core Features & Use Cases - SDTM Domain Validation: Checks required fields for DM (Demographics), LB (Laboratory), and VS (Vital Signs) domains against CDISC specifications. - Missing Value Handling & Outlier Detection: Imputes missing values with mean, median, mode, or forward-fill strategies, and detects outliers using IQR, Z-score, or clinical domain thresholds (e.g., glucose 50-500 mg/dL). - Date Standardization & Audit Trail: Converts dates to ISO 8601 format and logs every cleaning action to a JSON report suitable for 21 CFR Part 11 compliance. - Use Case: A clinical data manager preparing an FDA NDA submission runs the cleaner on raw lab data with domain-specific thresholds, producing a flagged SDTM dataset and a documented cleaning report for Pinnacle 21 validation. ## Quick Start Clean my raw demographics CSV file into an SDTM-compliant DM dataset using median imputation and flag any outliers, then save the audit trail report.
